Reimagining Data Management with AI Agents at Truist Bank

I recently had the chance to sit down with Jonathan Bruce (JB), Field CTO at Alation, at the Gartner Data & Analytics Conference 2026 in Orlando. JB and I spent 20 minutes together talking through something I think about every single day: how AI agents are reshaping what data management actually looks like inside a large financial institution — and what that means for the people doing that work.

I've been in the data space for about 15 years now, coming up through financial services, with some accounting background mixed in. I like to say I'm a bit of a jack-of-all-trades who burrowed into data and never looked back. And right now, I genuinely believe we're at one of the most interesting inflection points I've seen in this field. Here’s what I’ve learned so far. 

From governance to innovation

Most large financial institutions begin the data governance and management journey due to regulatory scrutiny. That's just reality. In banking and financial services, compliance isn't optional, and for good reason. Managing metadata, understanding data quality, mapping lineage — those things initially get done because we must do them to keep our business compliant and customers protected. That's the stick.

But what I try to help people understand is that compliance is the foundation, not the end-goal. Building out governance processes helps organizations understand their data, know where it lives, know how it moves — and as you progress, we get to do something much more interesting with it.

JB puts it well: you have to eat your vegetables first. And I think that's exactly right. Get the fundamentals right, and then you can start going after the real opportunities, or the carrot, as he calls it.

For me, the carrot is helping the business solve the problems that actually matter. I'll give you a concrete example from a previous job I held. Our marketing team came to us frustrated and wanted to improve the ROI on a campaign, and they believed they had a data problem. So we went into our data catalog, found the data they needed, identified the right steward for them to talk to, and helped put the right data-sharing agreement in place. The result? We improved that campaign's ROI by 20%.

That is real business value. That came directly from having our data house in order. And that organized house is a result of disciplined data governance.

AI agents extend your team

When people hear "AI agents" in the context of data management, I think there's sometimes a fear that it's a headcount conversation. I want to be clear: that's not how I think about it, or think it will be in practice as AI agents are built.

I see the opportunity to have AI agents automate lower-level tasks, allowing people to focus on higher-level tasks that produce business value. Why do I need someone to manually enter descriptions of data when an agent can do that? Why do I need a person to hand-code data quality rules when an agent can handle it? Those tasks exist right now; they take real time and take time away from work that genuinely requires human judgment.

What I want my technical data stewards and data analysts to do is the higher-level work. Analyzing and resolving data issues to build trust in data needed to solve business problems. Reporting back to the business on what's really happening. Driving toward resolution. That's where the human mind adds irreplaceable value, not in the mechanical tasks we've been asking people to grind through.

I think the sweet spot we're moving toward looks something like this: you identify a business outcome, you identify the data products supporting it, and you declare your standards for your metadata, data quality, and lineage, with agents continuously enforcing your standards and humans checking in at the right moments. The team isn't gone. They're just operating at a completely different level. AI can remove the drudgery and free up my team for much more rewarding work.

The metrics that will tell us we're getting it right

JB pushed me on this during our conversation, and it's a fair question: what does success actually look like when you have a mix of human and AI workers contributing to data management?

For me, the answer comes back to throughput and volume. Right now, data governance programs are focused on our most critical data: critical data elements, critical processes. But the real ambition is to expand: to govern and actively monitor a much broader data footprint across the enterprise, without having to proportionally grow the team to do it. That governed data is useful fuel for our future AI models. How do we get there?

If we can do more work with the same people — or the same work while those same people are solving harder problems — that's meaningful progress. We're not trying to reduce the team. We're trying to redeploy their talent where it creates more value.

And increasingly, I measure success by outcomes, not outputs. What I really want to know is: did we help a business team make a better decision? Did we unlock data that improved a customer experience? Did we cut time out of a process that was frustrating our mortgage teams? Those are the outcomes that justify the investment.

Data teams are becoming storytellers — and that's a good thing

Here's something I've been thinking about a lot. As agents absorb more of the mechanical work, data management professionals will have to develop a skill set that, frankly, our field hasn't always prioritized: communication.

Think about how much time we currently spend just identifying and chasing down the people responsible for critical data. Getting them to fill things in. Running down the lineage manually. That's time we're not spending talking to our business partners about why that data matters, what it means, and how it connects to the outcomes they care about.

I want my team to be good storytellers. I want them in the room with the marketing department, the mortgage team, and the data scientists, helping them use the data to drive action. That's a completely different role than what many data professionals have been asked to play historically — and it's a more valuable one.

My vision is that every analyst, every data scientist, every data modeler and engineer — not just the specialists — can understand and speak to their data. They know where to find what they need. They understand how to use it responsibly. They're empowered to build agents and solve problems. The knowledge of our data needs to be in the hands of our business.

It always comes back to solving the business problem and driving business outcomes. That's the north star.

Responsible innovation in the age of AI

My view on deploying AI at scale, particularly in a regulated industry, is straightforward: human oversight isn't optional. As agents take on more work, quality control becomes more important, not less. We still have to be able to look a regulator in the eye and say, we know what our agents are doing, we know why, and we have humans accountable for those outcomes. That accountability doesn't go away just because a machine is doing more of the work.

On the regulatory side, I think the honest answer is that the regulations are still catching up. We're all learning — practitioners and regulators alike. But we operate from a posture of getting ahead of that than waiting to be told what's required.

The future of AI in banking

If you ask me where the puck is heading over the next year or two, here's what I see:

The scope of what we govern is going to expand dramatically. We've been very focused on critical data elements, and rightfully so — but AI is going to let us cast a much wider net. More data, actively monitored, with the same teams.

The skills people need are going to shift. Less time running queries in the background, more time engaging with the business as real partners. Data people are going to have to get comfortable in rooms they haven't always been in, having conversations they haven't always been asked to have.

And the barriers between "data people" and "business people" will keep coming down. JB mentioned a CEO he'd heard about who is writing his own agents in Alation. That’s awesome! That kind of empowerment — where leaders at every level can engage with their data directly — is exactly where we should be heading.

AI isn't going to replace the people on my team. What it's going to do is finally give them the time and space to do the things they've always told me they needed to do, but couldn't get to because the day-to-day was too consuming. That's the transformation I'm here to lead.

The bottom line is this: data governance has always had the potential to be a genuine competitive advantage. For a long time, the work required to unlock that potential was just too manual, too slow, and too resource-intensive to scale. AI agents are changing that equation. The industry is still in the early stages, but the direction is clear, and the results we're already seeing make me genuinely optimistic about where we're headed.

Eat your vegetables. Build the foundation. Then use AI to do things you never had the bandwidth to do before. That's the playbook.

Gary Dugan is SVP, Head of Data Management Execution at Truist Bank. This conversation originally took place at the Gartner Data & Analytics Conference 2026 in Orlando, in a session with Jonathan Bruce, Field CTO at Alation.